摘 要 逆物流的導入，能有效降低產品回收的成本，提升顧客滿意程度，以增加利潤；近年來，逆物流的發展漸為各企業所重視；企業常藉由正向物流的正常營運，再全盤性的導入逆物流。在物流的過程中，實體運輸扮演著不可或缺的角色（運輸成本佔了供應鏈總成本的52﹪），正、逆向貨品的順利配送、撿收實有賴運輸服務的妥善規劃。本研究因此以物流業者在特定時窗內完成顧客所需配送、撿收服務為目的，同時考量正、逆向貨品的流動，即以配送為主的車輛繞徑問題同時完成撿收貨品的需求。傳統的車輛途程問題，多設定單一目標（總行駛距離最短為主），但實務上，企業的營運目標多非單一，而是多方面考量，因此本研究同時考量企業營運成本及顧客服務水準，以總行駛距離最短及總違反時窗成本最小為雙目標考量，利用權重加以整合。在本研究中，提出以運送車輛卸貨後滕出之空間進行取貨，故僅需運送車輛內剩餘空間大於該點顧客的取貨需求量，即行撿收，與以往傳統方法需運送車輛內空出一定比例之後，才行撿收的設計不同，本研究並以軟性時窗懲罰成本設計（即隨違反時間的增加，懲罰倍增，以呈現實務上，顧客對於時間的價值性），所羅門標竿題庫(Solomon Benchmark Testing Bank) 經修改後，隨機產生撿收需求，作為基因演算法測試基礎，自行撰寫程式加以驗證。測試結果發現隨著撿收需求量的增加，總成本（含總行駛距離成本以及總違反時窗成本）隨之微幅增加，取貨量約為送貨量的一半時，最多僅增加6.44%的成本，由此可驗證本研究所提出之以車輛卸貨後之空間實行取貨服務，確實能在成本微幅增加內，有效的提高車輛使用率，降低運送車輛回程的空車容量。 Abstract The introduction of reverse logistics may significantly reduce the cost of returned merchandise, improve the customer’s satisfaction, and therefore increase enterprise profit. Recently, most of enterprises pay their great attention to the inclusion of reverse logistics to proceed hand in hand with already regularly-operated forward logistics. Transportation of physical distribution plays a critical role in logistics. In general, transportation cost is the majority of the total cost in supply chain, fifty-two percent approximately. The principal objective of the study is therefore to concurrently accomplish the delivery and pickup commodities at customer’s specified time window, i.e. to perform forward and reverse logistics at the identical trip. It may be classified as a vehicle routing problem with soft time window for simultaneous commodity delivery and pickup. The traditional vehicle routing problem is mostly considered the single object, the total distance, however the operation of companies is based on multiple objects rather than a single object. Therefore this study is to integrate the shortest distance and the fewest penalty cost by the weighting method in terms of both the cost and the level of service. This study suggested that pickup during the delivery only requires the rest capacity enough for the goods. There is no such constraint that pickup only can be performed when the certain capacity of the car is left. We solved the modified Solomon benchmark with random pickup demand by using Genetic Algorithms with soft time window (The penalty is dramatically increased by increasing the time of being late and waiting to emphasize the value of time for customers). In addition, we program a computer to demonstrate the accuracy of this idea. The results showed that total cost was slightly increased by increasing the demand of pickup and the cost was increased maximally about 6.44% when the quantity of pickup was about half of that of delivery. As a result, the idea, pickup during the delivery, actually improved the efficiency of the vehicle usage and reduced the waste of the capacity of returned vehicles.
